1 Statics and dynamics: some elementary concets Dynamics is the study of the movement through time of variables such as heartbeat, temerature, secies oulation, voltage, roduction, emloyment, rices and so forth. This is often achieved by means of equations linking the values of variables at different, uniformly saced instants of time, i.e., difference equations, or by systems relating the values of variables to their time derivatives, i.e., ordinary differential equations. Dynamical henomena can also be investigated by other tyes of mathematical reresentations, such as artial differential equations, lattice mas or cellular automata. In this book, however, we shall concentrate on the study of systems of difference and differential equations and their dynamical behaviour. In the following chaters we shall occasionally use models drawn from economics to illustrate the main concets and methods. However, in general, the mathematical roerties of equations will be discussed indeendently of their alications. 1.1 A static roblem To rovide a first, broad idea of the roblems osed by dynamic vis-à-vis static analysis, we shall now introduce an elementary model that could be labelled as suly-demand-rice interaction in a single market. Our model considers the quantities sulied and demanded of a single good, defined as functions of a single variable, its rice,. In economic arlance, this would be called artial analysis since the effect of rices and quantities determined in the markets of all other goods is neglected. It is assumed that the demand function D() is decreasing in (the lower the rice, the greater the amount that eole wish to buy), while the suly function S() is increasing in (the higher the rice, the greater the amount that eole wish to suly). 1
2 Statics and dynamics: some elementary concets D,S D D( ) = S( ) D S Fig. 1.1 The static artial equilibrium model For examle, in the simler, linear case, we have: D() =a b S() = m + s (1.1) and a, b, m and s are ositive constants. Only nonnegative values of these variables are economically meaningful, thus we only consider D, S, 0. The economic equilibrium condition requires that the market of the good clears, that is demand equals suly, namely: D() =S() (1.2) or a b = m + s. static solution Mathematically, the solution to our roblem is the value of the variable that solves (1.2) (in this articular case, a linear equation). Solving (1.2) for we find: = a + m b + s where is usually called the equilibrium rice (see figure 1.1). 1 We call the roblem static since no reference is made to time or, if you refer, 1 The demand curve D in figure 1.1 is rovided to make the oint that, with no further constraints on arameter values, the equilibrium rice could imly negative equilibrium quantities of suly and demand. To eliminate this ossibility we further assume that 0 <m/s a/b, as is the case for the demand curve D.
1.2 A discrete-time dynamic roblem 3 everything haens at the same time. Notice that, even though the static model allows us to find the equilibrium rice of the good, it tells us nothing about what haens if the actual rice is different from its equilibrium value. 1.2 A discrete-time dynamic roblem The introduction of dynamics into the model requires that we relace the equilibrium condition (1.2) with some hyothesis concerning the behaviour of the system off-equilibrium, i.e., when demand and suly are not equal. For this urose, we assume the most obvious mechanism of rice adjustment: over a certain interval of time, the rice increases or decreases in roortion to the excess of demand over suly, (D S) (for short, excess demand). Of course, excess demand can be a ositive or a negative quantity. Unless the adjustment is assumed to be instantaneous, rices must now be dated and n denotes the rice of the good at time n, time being measured at equal intervals of length h. Formally, we have n+h = n + hθ[d( n ) S( n )]. (1.3) Since h is the eriod of time over which the adjustment takes lace, θ can be taken as a measure of the seed of rice resonse to excess demand. For simlicity, let us choose h =1,θ = 1. Then we have, making use of the demand and suly functions (1.1), n+1 = a + m +(1 b s) n. (1.4) In general, a solution of (1.4) is a function of time (n) (with n taking discrete, integer values) that satisfies (1.4). 2 dynamic solution To obtain the full dynamic solution of (1.4), we begin by setting α = a + m, β =(1 b s) to obtain n+1 = α + β n. (1.5) To solve (1.5), we first set it in a canonical form, with all time-referenced terms of the variable on the left hand side (LHS), and all constants on the right hand side (RHS), thus: Then we roceed in stes as follows. n+1 β n = α. (1.6) 2 We use the forms n and (n) interchangeably, choosing the latter whenever we refer to emhasise that is a function of n.
4 Statics and dynamics: some elementary concets ste 1 We solve the homogeneous equation, which is formed by setting the RHS of (1.6) equal to 0, namely: n+1 β n =0. (1.7) It is easy to see that a function of time (n) satisfying (1.7) is (n) =Cβ n, with C an arbitrary constant. Indeed, substituting in (1.7), we have Cβ n+1 βcβ n = Cβ n+1 Cβ n+1 =0. ste 2 We find a articular solution of (1.6), assuming that it has a form similar to the RHS in the general form. Since the latter is a constant, set (n) = k, k a constant, and substitute it into (1.6), obtaining so that k = k βk = α α 1 β = a + m b + s = again! It follows that the (n) = is a solution to (1.6) and the constant (or stationary) solution of the dynamic roblem is simly the solution of the static roblem of section 1.1. ste 3 Since (1.6) is linear, the sum of the homogeneous and the articular solution is again a solution, 3 called the general solution. This can be written as (n) = + Cβ n. (1.8) The arbitrary constant C can now be exressed in terms of the initial condition. Putting (0) 0, and solving (1.8) for C we have 0 = + Cβ 0 = + C whence C = 0, that is, the difference between the initial and equilibrium values of. The general solution can now be re-written as (n) = +( 0 )β n. (1.9) Letting n take integer values 1, 2,..., from (1.9) we can generate a sequence of values of, a history of that variable (and consequently, a history of quantities demanded and sulied at the various rices), once its value at any arbitrary instant of time is given. Notice that, since the function n+1 = 3 This is called the suerosition rincile and is discussed in detail in chater 2 section 2.1.
1.2 A discrete-time dynamic roblem 5 f( n )isinvertible, i.e., the function f 1 is well defined, n 1 = f 1 ( n ) also describes the ast history of. The value of at each instant of time is equal to the sum of the equilibrium value (the solution to the static roblem which is also the articular, stationary solution) and the initial disequilibrium ( 0 ), amlified or damened by a factor β n. There are therefore two basic cases: (i) β > 1. Any nonzero deviation from equilibrium is amlified in time, the equilibrium solution is unstable and as n +, n asymtotically tends to + or. (ii) β < 1. Any nonzero deviation is asymtotically reduced to zero, n as n + and the equilibrium solution is consequently stable. First-order, discrete-time equations (where the order is determined as the difference between the extreme time indices) can also have fluctuating behaviour, called imroer oscillations, 4 owing to the fact that if β<0, β n will be ositive or negative according to whether n is even or odd. Thus the sign of the adjusting comonent of the solution, the second term of the RHS of (1.9), oscillates accordingly. Imroer oscillations are damened if β> 1 and exlosive if β< 1. In figure 1.2 we have two reresentations of the motion of through time. In figure 1.2(a) we have a line defined by the solution equation (1.5), and the bisector assing through the origin which satisfies the equation n+1 = n. The intersection of the two lines corresonds to the constant, equilibrium solution. To describe the off-equilibrium dynamics of, we start on the abscissa from an initial value 0. To find 1, we move vertically to the solution line and sidewise horizontally to the ordinate. To find 2, we first reflect the value of 1 by moving horizontally to the bisector and then vertically to the abscissa. From the oint 1, we reeat the rocedure roosed for 0 (u to the solution line, left to the ordinate), and so on and so forth. The rocedure can be simlified by omitting the intermediate ste and simly moving u to the solution line and sidewise to the bisector, u again, and so on, as indicated in figure 1.2(a). It is obvious that for β < 1, at each iteration of the rocedure the initial deviation from equilibrium is diminished again, see figure 1.2(b). For examle, if β = 0.7, we have β 2 =0.49, β 3 =0.34,...,β 10 0.03,...) and the equilibrium solution is aroached asymtotically. The reader will notice that stability of the system and the ossibility 4 The term imroer refers to the fact that in this case oscillations of variables have a kinky form that does not roerly describe the smoother us and downs of real variables. We discuss roer oscillations in chater 3.
6 Statics and dynamics: some elementary concets n+1 0 (a) 0 1 n (b) 1 2 3 4 5 6 7 time Fig. 1.2 Convergence to in the discrete-time artial equilibrium model of oscillatory behaviour deends entirely on β and therefore on the two arameters b and s, these latter denoting resectively the sloes of the demand and suly curves. The other two arameters of the system, a and m, determine α and consequently they affect only the equilibrium value. We can therefore comletely describe the dynamic characteristics of the solution (1.9) over the arameter sace (b, s). The boundary between stable and unstable behaviour is given by β = 1, and convergence to equilibrium is guaranteed for 1 <β<1 2 >b+ s>0. The assumtions on the demand and suly functions imly that b, s > 0. Therefore, the stability condition is (b + s) < 2, the stability boundary is the line (b + s) = 2, as reresented in figure 1.3. Next, we define the curve β =1 (b + s) = 0, searating the zone of monotonic behaviour from that of imroer oscillations, which is also reresented in figure 1.3. Three zones are labelled according to the different tyes of dynamic behaviour, namely: convergent and monotonic; convergent and oscillatory; divergent and oscillatory. Since b, s > 0, we never have the case β>1, corresonding to divergent, nonoscillatory behaviour. If β > 1 any initial difference ( 0 ) is amlified at each ste. In this model, we can have β > 1 if and only if β< 1. Instability, then, is due to overshooting. Any time the actual rice is, say, too low and there is ositive excess demand, the adjustment mechanism generates a change in the rice in the right direction (the rice rises) but the change is too large.
1.3 A continuous-time dynamic roblem 7 s 2 Convergent, monotonic Convergent, oscillatory Divergent, oscillatory 1 1 2 b Fig. 1.3 Parameter sace for the discrete-time artial equilibrium model After the correction, the new rice is too high (negative excess demand) and the discreancy from equilibrium is larger than before. A second adjustment follows, leading to another rice that is far too low, and so on. We leave further study of this case to the exercises at the end of this chater. 1.3 A continuous-time dynamic roblem We now discuss our simle dynamical model in a continuous-time setting. Let us consider, again, the rice adjustment equation (1.3) (with θ = 1, h>0) and let us adot the notation (n) so that (n + h) =(n)+h (D[(n)] S[(n)]). Dividing this equation throughout by h, we obtain (n + h) (n) = D[(n)] S[(n)] h whence, taking the limit of the LHS as h 0, and recalling the definition of a derivative, we can write d(n) dn = D[(n)] S[(n)]. Taking the interval h to zero is tantamount to ostulating that time is a continuous variable. To signal that time is being modelled differently we substitute the time variable n Z with t R and denote the value of at time t simly by, using the extended form (t) when we want to emhasise that rice is a function of time. We also make use of the efficient Newtonian
8 Statics and dynamics: some elementary concets notation dx(t)/dt = ẋ to write the rice adjustment mechanism as d dt =ṗ = D() S() =(a + m) (b + s). (1.10) Equation (1.10) is an ordinary differential equation relating the values of the variable at a given time t to its first derivative with resect to time at the same moment. It is ordinary because the solution (t) is a function of a single indeendent variable, time. Partial differential equations, whose solutions are functions of more than one indeendent variable, will not be treated in this book, and when we refer to differential equations we mean ordinary differential equations. dynamic solution The dynamic roblem is once again that of finding a function of time (t) such that (1.10) is satisfied for an arbitrary initial condition (0) 0. As in the discrete-time case, we begin by setting the equation in canonical form, with all terms involving the variable or its time derivatives on the LHS, and all constants or functions of time (if they exist) on the RHS, thus Then we roceed in stes as follows. ṗ +(b + s) = a + m. (1.11) ste 1 We solve the homogeneous equation, formed by setting the RHS of (1.11) equal to 0, and obtain ṗ +(b + s) = 0 or ṗ = (b + s). (1.12) If we now integrate (1.12) by searating variables, we have whence d = (b + s) dt ln (t) = (b + s)t + A where A is an arbitrary integration constant. Taking now the antilogarithm of both sides and setting e A = C, we obtain (t) =Ce (b+s)t.
1.3 A continuous-time dynamic roblem 9 ste 2 We look for a articular solution to the nonhomogeneous equation (1.11). The RHS is a constant so we try = k, k a constant and consequently ṗ = 0. Therefore, we have whence ṗ =0=(a + m) (b + s)k k = a + m b + s =. Once again the solution to the static roblem turns out to be a secial (stationary) solution to the corresonding dynamic roblem. ste 3 Since (1.12) is linear, the general solution can be found by summing the articular solution and the solution to the homogeneous equation, thus (t) = + Ce (b+s)t. Solving for C in terms of the initial condition, we find (0) 0 = + C and C =( 0 ). Finally, the comlete solution to (1.10) in terms of time, arameters, initial and equilibrium values is (t) = +( 0 )e (b+s)t. (1.13) As in the discrete case, the solution (1.13) can be interreted as the sum of the equilibrium value and the initial deviation of the rice variable from equilibrium, amlified or damened by the term e (b+s)t. Notice that in the continuous-time case, a solution to a differential equation ṗ = f() always determines both the future and the ast history of the variable, indeendently of whether the function f is invertible or not. In general, we can have two main cases, namely: (i) (b + s) > 0 Deviations from equilibrium tend asymtotically to zero as t +. (ii) (b + s) < 0 Deviations become indefinitely large as t + (or, equivalently, deviations tend to zero as t ). Given the assumtions on the demand and suly functions, and therefore on b and s, the exlosive case is excluded for this model. If the initial rice is below its equilibrium value, the adjustment rocess ensures that the rice increases towards it, if the initial rice is above equilibrium, the rice declines to it. (There can be no overshooting in the continuous-time case.) In a manner analogous to the rocedure for difference equations, the equilibria
10 Statics and dynamics: some elementary concets ṗ 30 20 10 (a) 0 (b) 5 10 15 time Fig. 1.4 The continuous-time artial equilibrium model of differential equations can be determined grahically in the lane (, ṗ) as suggested in figure 1.4(a). Equilibria are found at oints of intersection of the line defined by (1.10) and the abscissa, where ṗ = 0. Convergence to equilibrium from an initial value different from the equilibrium value is shown in figure 1.4(b). Is convergence likely for more general economic models of rice adjustment, where other goods and income as well as substitution effects are taken into consideration? A comrehensive discussion of these and other related microeconomic issues is out of the question in this book. However, in the aendixes to chater 3, which are devoted to a more systematic study of stability in economic models, we shall take u again the question of convergence to or divergence from economic equilibrium. We would like to emhasise once again the difference between the discretetime and the continuous-time formulation of a seemingly identical roblem, reresented by the two equations n+1 n =(a + m) (b + s) n (1.4) ṗ =(a + m) (b + s). (1.10) Whereas in the latter case (b+s) > 0 is a sufficient (and necessary) condition for convergence to equilibrium, stability of (1.4) requires that 0 < (b+s) < 2, a tighter condition. This simle fact should make the reader aware that a naive translation of a model from discrete to continuous time or vice versa may have unsusected consequences for the dynamical behaviour of the solutions.